Athelas To Pioneer in Agentic Practice Management Spaces
-
Click here to listen to this article - Share via
- Athelas released software that lets clinic operations staff assign billing, scheduling and patient messaging work to AI agents. The agents complete a task rather than recommend one.
- Athelas describes the product as the first of its kind. Two competing billing software vendors announced agentic tools earlier in the same year.
- Physicians in the one randomized trial of a neighboring AI tool reported that its output occasionally contained clinically significant errors.
EDITOR’S NOTE: This article was updated on September 15, 2026 after a post-publication review. The original version described the company’s claim to be first in its category without giving the basis for it, and characterized athenahealth as a billing software vendor when athenaOne is a practice management and records suite. The updated version gives the competing announcement dates and, where the companies stated them, which features were generally available rather than in testing. The original also faulted a business associate agreement for not addressing AI oversight; that agreement governs patient data handling and is not the document that defines product limits. That passage has been rewritten around the general terms of agreement, which is the document that does, and around the company’s account of its control model. A summary point citing error rates in unrelated AI scribe products has been removed from Key Facts. The research on ambient AI documentation remains, with its relevance stated more precisely and the company’s objection to the comparison included.
Table of Contents
- What Athelas’s Agentic Practice Management Suite Does
- Why Clinic Administration Is the Target
- What Customer Practices Report
- Where Adoption Stalls
- Research Gap
- Closing Thoughts
Athelas released Practice Manager on August 5, 2026, and called it the industry’s first suite in a category the company named agentic practice management [1]. The product lets clinic operations teams hand billing, scheduling and patient communication work to software agents, programs that carry a task through to the end rather than recommend a next step.
The superlative is the company’s own, and it is not confined to the announcement. Athelas’s product page for Practice Manager is headlined “The World’s First Agentic Practice Management Suite” [11].
What the company points to is the combination. The announcement describes agents running billing, scheduling and patient communication inside one suite, generally available at launch rather than staged in over following quarters [1, 11].
Competitors announced parts of that ground earlier. Waystar announced agentic capabilities across its AltitudeAI revenue cycle platform in January 2026 [8]. athenahealth announced AI across practice and revenue cycle work in athenaOne in October 2025, agentic text and voice tools for patient communication in February 2026, and more than 80 AI revenue cycle features in June 2026 [9, 12, 13].
What was generally available is the narrower question, and it is the one the claim turns on. athenahealth’s October announcement listed two features as live in athenaOne, both on the billing side, with waitlist scheduling and its payer portal agents in alpha testing or earlier [13]. In February its patient texting assistant was in alpha, and the voice front desk assistant and the patient portal assistant were slated for the second half of the year [12].
What Athelas’s Agentic Practice Management Suite Does
Practice Manager sits alongside Athelas’s existing products: the Air EHR, ambient documentation, outsourced revenue cycle management services and voice agents.
It gives staff one place to see revenue, trace a problem to the claims behind it, and correct denials in bulk.
Athelas says three functions run with limited human input [1]. Rules written in plain English check claims for errors before submission, an AI copilot drafts corrections for denied claims, and payment posting runs on its own from bank deposits through to the explanation of benefits, the insurer’s line-by-line account of what it paid and what it declined.
Hersh Solanki, chief business officer at Athelas, told LA Times Studios in September 2026 that the constraint holding agents back is data. Clinics keep records across an EHR, a clearinghouse, payer portals and a front desk system, and a person moves between them by switching screens all day.
“An agent can only finish a task when it sees the whole picture,” Solanki said. “A denial correction, for example, requires the clinical note, the eligibility check, the payer history, and the remittance in one place.”
Athelas is owned by Commure and backed by General Catalyst. The company says the combined platform now covers about 7% of US healthcare appointments and processes billions of dollars in claims each year [1].
Athelas does not publish pricing for Practice Manager. Its product page lists no price and no licensing model, and directs buyers to request a demo [11]. Waystar’s site does the same, so that is a category norm rather than one company’s choice.
The competing products overlap on function, though not on scope. Waystar’s AltitudeAI works on the revenue cycle, covering prior authorization justification, denial prevention and automated clinical appeals, and the company says its agents work with minimal human intervention [8]. athenaOne is a practice management and records suite, and athenahealth says its Expanded Denial Resolution Automation generates claim corrections for coding denials now, with automated resubmission and payer portal appeals due through 2026 [9].
Athelas argues that running agents across billing, scheduling and patient messaging at once is a different product than adding agents to one of them. The distinction is real, and it is also the distinction its “first” claim depends on.
Why Clinic Administration Is the Target
The administrative burden on US medical practice is large and well measured. A time and motion study of 57 physicians across four specialties found they spent 49.2% of the office day on EHR and desk work and 27.0% in direct contact with patients, then 1 to 2 more hours on clerical work each night [2]. US insurers and providers spent $812 billion on administration in 2017, or 34.2% of national health expenditures, against 17.0% in Canada [3].
Prior authorization is the sharpest version of the problem. Practices complete an average of 40 prior authorization requests per physician each week and spend 13 hours of staff and physician time on them, and 94% of physicians surveyed by the American Medical Association said it contributes to burnout [4]. Prior authorization and denial work is the share of that burden Practice Manager claims to automate, the same staffing math behind AI tools deployed against specialist shortages elsewhere in medicine.
What Customer Practices Report
Matt Ernst, PT, MPT, OCS, COMT, CIDN, a partner and vice president of operations at Oxford Physical Therapy Centers, an 11-location group across Ohio and northern Kentucky, told LA Times Studios in September 2026 that the savings showed up in staffing decisions.
“We have been able to grant requests from staff for shortened schedules and combine some front desk and billing roles into hybrid dual roles,” Ernst said. “That would not have been possible without the efficiencies gained utilizing the Athelas platform.”
Ernst said his group used the rules engine to cut prior authorization denials 13% in one month. The rules, he said, combined what insurance verification, the front desk and billing each saw separately, then directed providers to confirm authorization before the claim went out.
Athelas reports larger figures from other customers. The company says Lattimore Physical Therapy cut billing costs 75% while scaling to 35 additional locations, and that Oxford PT recorded its largest revenue month in company history on the platform [1].
Solanki told LA Times Studios that Cape Cod Rehab, a physical therapy provider in Massachusetts, reports 6 to 7 hours returned to staff each week, and that Raleigh Orthopaedic, a group of more than 140 providers, cut its billing team’s pending work items 81.5% in a single week through automated review of authorization denials. Athelas has not published the methodology behind either figure.
Where Adoption Stalls
Past the data problem, Solanki said the obstacle inside a single clinic is habit. Front desk staff have verified insurance the same way for a decade and billers work every denial by hand, and moving to AI agents asks them to trust output they once produced themselves.
Solanki described a staged rollout. Agents draft, staff approve, every action stays visible, and the team writes the operating rules in ordinary English. The friction is not unique to billing offices: specialty societies have hesitated to endorse AI tools in clinical settings for the same reason.
Multi-location practices hit a second obstacle. “A group with ten clinics usually runs 10 versions of the same process,” Solanki said, and automation forces leadership onto one standard, because an agent needs an agreed process before it can run one.
Chris Murray, a managing director at Huron Consulting Group and an analyst outside the vendor set, told TechTarget that most vendors and providers are still building the data governance and security foundations that end-to-end automation requires [8]. “This is where rubber hits the road, and it’s a lot more challenging than all those releases,” Murray said.
That oversight is contractual as well as technical. Commure publishes a business associate agreement committing it to the HIPAA Security Rule’s safeguards for electronic health information and binding any subcontractor to the same terms [10]. Its trust center lists SOC 2 Type II auditing and says its products and workflows meet HIPAA and HITECH standards [14].
Those documents govern patient data. Where the product’s own limits are written down is the general terms of agreement, which took effect in November 2025 and is public [15].
That document is direct about what the software is for. Services with AI features “may be incomplete, inaccurate, or biased,” it says, and those features “are intended to assist, not replace, professional judgment” [15]. It places responsibility for verifying accuracy on the customer before they rely on what the software produces.
Solanki said the question worth asking is a different one. “The appropriate question is therefore not whether the agreement uses the term AI, but what controls govern the system’s actions, what permissions customers provide, and how those actions are monitored and audited,” he told LA Times Studios.
His answer is that the practice sets them. Practice Manager runs inside customer-defined workflows and permissions, he said, the rules engine is written by the customer in plain English, and every agent action is logged, auditable and visible in the platform.
He was also explicit about what the company does not claim. Athelas does not represent that a person reviews every claim before an agent submits it, and Solanki said that is the point rather than a gap: software executing defined, repetitive administrative work inside rules the customer wrote is different from an unrestricted system making arbitrary decisions.
Read together, the terms and the control model put the same thing in two places. The practice writes the rules, the practice checks the output, and the log is what makes checking possible. No outside party has audited how those controls behave in production.
Research Gap
No peer-reviewed study has measured what agent-run billing software does to claim accuracy, denial rates or clinic staffing. That is true of Athelas and of every competitor named here.
Athelas has not published error rates for Practice Manager and did not provide performance data for this article.
Solanki objected to the comparison that follows, and the objection is worth stating plainly. Practice Manager is purpose-built with its own architecture, training data and controls, he said, and shares nothing with the products below “beyond the fact that all of them can be described with the words artificial intelligence. That is a category label, not a shared codebase.”
What follows is not a measurement of Practice Manager. It is the nearest category anyone has measured independently, ambient AI documentation, and it is here because its results came in below what vendors reported.
A three-group randomized trial at UCLA assigned 238 outpatient physicians to one of two ambient AI scribes or usual care. Nabla users cut time-in-note by 9.5% against the control group, DAX users showed no significant change, and physicians using either product reported clinically significant inaccuracies “occasionally” on a five-point scale [6]. A smaller Cleveland Clinic study of 10 primary care physicians found that the only measure to change significantly over 4 weeks was characters typed, which fell by 15,398 [7].
Ambient AI documentation is not the only clinical-AI category where gains that look large in a vendor’s own numbers shrink under independent measurement. Whether agent-run billing follows that pattern is unknown, and will stay unknown until someone measures it.
The scale of what is left to automate has been quantified. CAQH, which tracks administrative transactions across plans covering 63% of insured lives, put the remaining savings from fully automating manual and partly manual transactions at $21 billion [5].
Closing Thoughts
Practice Manager is a shipping product with paying customers and reported results, and the administrative problem it addresses is documented across the peer-reviewed literature. The claim to being first rests on a category boundary the company drew itself, which is worth knowing rather than damning.
What stays unmeasured is whether agent-run billing holds accuracy at scale, and whether the hours it returns go to patient care or to absorbing more volume on the same staff. Ernst’s answer, that his group used the savings to shorten schedules and merge roles, is one 11-location practice reporting on itself rather than a finding.
Readers weighing a clinical or coverage decision should consult their own clinician.
References
1. Commure. (2026, August 5). Athelas launches Practice Manager, the world’s first agentic practice management suite [Press release]. GlobeNewswire. https://www.globenewswire.com/news-release/2026/08/05/3339372/0/en/athelas-launches-practice-manager-the-world-s-first-agentic-practice-management-suite.html
2. Sinsky, C., Colligan, L., Li, L., Prgomet, M., Reynolds, S., Goeders, L., Westbrook, J., Tutty, M., & Blike, G. (2016). Allocation of physician time in ambulatory practice: A time and motion study in 4 specialties. Annals of Internal Medicine, 165(11), 753-760. https://doi.org/10.7326/M16-0961
3. Himmelstein, D. U., Campbell, T., & Woolhandler, S. (2020). Health care administrative costs in the United States and Canada, 2017. Annals of Internal Medicine, 172(2), 134-142. https://doi.org/10.7326/M19-2818
4. American Medical Association. (2026). 2025 AMA prior authorization physician survey. https://www.ama-assn.org/system/files/prior-authorization-survey.pdf
5. CAQH. (2026, February 19). 2025 CAQH Index shows U.S. healthcare avoided $258 billion and accelerated automation, interoperability and AI adoption. DataSpring. https://www.dataspring.com/blog/2025-caqh-index-shows-u.s.-healthcare-avoided-258-billion-and-accelerated-automation-interoperability-and-ai-adoption
6. Lukac, P. J., Turner, W., Vangala, S., Chin, A. T., Khalili, J., Shih, Y. T., Sarkisian, C., Cheng, E. M., & Mafi, J. N. (2025). Ambient AI scribes in clinical practice: A randomized trial. NEJM AI, 2(12). https://doi.org/10.1056/aioa2501000
7. Alpert, J. M., Saper, R., Boose, E., Ruff, J., Hopkins, K., Gaskins, D., Guo, N., Schneider, J. P., Giffi Scibona, A., Gutierrez, J., & Rothberg, M. B. (2025). Evaluating an artificial intelligence scribe for clinical documentation. Digital Health, 11, 20552076251395588. https://doi.org/10.1177/20552076251395588
8. LaPointe, J. (2026, January 13). Waystar adds agentic AI in move to “autonomous” revenue cycle. TechTarget. https://www.techtarget.com/revcyclemanagement/news/366637224/Waystar-adds-agentic-AI-in-move-to-autonomous-revenue-cycle
9. athenahealth. (2026, June 3). athenahealth raises the bar on revenue cycle management performance, with 80+ AI-native features designed to eliminate manual revenue cycle work [Press release]. https://www.athenahealth.com/about/press-releases/ai-revenue-cycle-management-features
10. Commure. (2025, June 24). Business associate agreement. Retrieved September 8, 2026, from https://www.commure.com/legal/business-associate-agreement/current
11. Athelas. (n.d.). Practice Manager. Retrieved September 8, 2026, from https://athelas.com/practice-manager
12. athenahealth. (2026, February 19). athenahealth launches agentic patient communication tools across its provider network [Press release]. Reprinted by HLTH. https://hlth.com/insights/news/athenahealth-introduces-agentic-ai-tools-to-modernize-patient-communication-across-its-network-2026-02-20
13. athenahealth. (2025, October 9). athenahealth introduces AI-native athenaOne practice and revenue cycle management capabilities designed to reduce practices’ administrative work by more than 50% [Press release]. Business Wire. https://markets.financialcontent.com/stocks/article/bizwire-2025-10-9-athenahealth-introduces-ai-native-athenaone-practice-and-revenue-cycle-management-capabilities-designed-to-reduce-practices-administrative-work-by-more-than-50
15. Commure. (2025, November 18). General terms of agreement. https://www.athelas.com/general-terms-of-agreement
14. Commure. (n.d.). Trust center. Retrieved September 14, 2026, from https://www.commure.com/trust-center
